Wanted to bump this thread to mention something I've been looking at for some time now.
One of the most common arguments (against advanced stats) I hear/read, once an individual has gotten to a point where they agree that a player spending more time in the offensive zone creating shots is better, is something along of the lines of : "Okay, but how can this data help a team improve/win?"
The first thing that pops into my head is player acquisition. Would a poor possession team improve by adding a possession-driving player? Perhaps. But trades are difficult and adding UFAs can only happen between seasons. What else?
For me, it's coaching, and specifically line combinations. Everyone knows that certain players player better with certain linemates. It doesn't take long to see the chemistry between Gaudreau/Hudler, or Giordano/Brodie. But can this chemistry be tracked statistically to be utilized when it may be less-than-obvious visually?
To illustrate, I looked at the combination of Giordano and Brodie over the past 1.5+ seasons. On the ice together, the Flames own 52.8% of the shot attempts, while most often starting in their own end, against the best players of the opposing team much of the time. On the ice with different partners, the Flames are under 47%. Gaudreau and Hudler are a similar story. This year, they are at 49% together but 44% with a different wing partner.
My thought was to find and exploit these numbers, in an attempt to optimize the lineup. The 12 forwards I used are the 12 in top total ice time, with Stajan (13th) replacing Granlund (10th) to match the current lineup. I attempted to only use tested combinations (minimum 50 minutes together) to get a decent sample size. If numbers for 2 linemates were similar, priority was given to a larger sample size. The reason I bring it up today is that tomorrow's projected lineup almost exactly mirrors mine.
Presenting,
The Flames' Corsi-Lineup
Gaudreau-Jooris-Hudler
Byron-Monahan-Colborne
Bouma-Backlund-Jones
Glencross-Stajan-Raymond
Comparing to tomorrow's projection, via SureLoss...
Quote:
Originally Posted by sureLoss
Lines:
Gaudreau-Monahan-Hudler
Bouma-Backlund-Jones
Glencross-Stajan-Raymond
Byron-Jooris-Colborne
|
Jooris and Monahan are flipped, but the rest is exactly the same. There was some variability with the wingers of lines 2 and 4 (in my lineup), but this was optimal. Assuming 60 minutes of 5v5, I would give lines 1 and 2 about 17 minutes, and 13 for lines 3 and 4. With the amount of special teams played, total ice time for each player could be more or less the same as their current averages.
Did we need 'advanced' stats to get here? Maybe not, as Hartley has pretty much reached the same conclusion as me. But it's certainly nice to see the numbers back up the eye test.
Calgarypuck Advanced Stat Summary
